Facilitating Action for Suicide Prevention by Learning Health Care Systems

Stability of Suicide Risk Prediction Models During Changes in Health Care Delivery

OBJECTIVE

The authors aimed to use health records data to examine how the accuracy of statistical models predicting self-harm or suicide changed between 2015 and 2019, as health systems implemented suicide prevention programs.

METHODS

Data from four large health systems were used to identify specialty mental health visits by patients ages ≥11 years, assess 311 potential predictors of self-harm (including demographic characteristics, historical risk factors, and index visit characteristics), and ascertain fatal or nonfatal self-harm events over 90 days after each visit. New prediction models were developed with logistic regression with LASSO (least absolute shrinkage and selection operator) in random samples of visits (65%) from each calendar year and were validated in the remaining portion of the sample (35%).

RESULTS

A model developed for visits from 2009 to mid-2015 showed similar classification performance and calibration accuracy in a new sample of about 13.1 million visits from late 2015 to 2019. Area under the receiver operating characteristic curve (AUC) ranged from 0.840 to 0.849 in the new sample, compared with 0.851 in the original sample. New models developed for each year for 2015-2019 had classification performance (AUC range 0.790-0.853), sensitivity, and positive predictive value similar to those of the previously developed model. Models selected similar predictors from 2015 to 2019, except for more frequent selection of depression questionnaire data in later years, when questionnaires were more frequently recorded.

CONCLUSIONS

A self-harm prediction model developed with 2009-2015 visit data performed similarly when applied to 2015-2019 visits. New models did not yield superior performance or identify different predictors.

Predicting risk of suicidal behavior from insurance claims data vs. linked data from insurance claims and electronic health records

PURPOSE

Observational studies assessing effects of medical products on suicidal behavior often rely on health record data to account for pre-existing risk. We assess whether high-dimensional models predicting suicide risk using data derived from insurance claims and electronic health records (EHRs) are superior to models using data from insurance claims alone.

METHODS

Data were from seven large health systems identified outpatient mental health visits by patients aged 11 or older between 1/1/2009 and 9/30/2017. Data for the 5 years prior to each visit identified potential predictors of suicidal behavior typically available from insurance claims (e.g., mental health diagnoses, procedure codes, medication dispensings) and additional potential predictors available from EHRs (self-reported race and ethnicity, responses to Patient Health Questionnaire or PHQ-9 depression questionnaires). Nonfatal self-harm events following each visit were identified from insurance claims data and fatal self-harm events were identified by linkage to state mortality records. Random forest models predicting nonfatal or fatal self-harm over 90 days following each visit were developed in a 70% random sample of visits and validated in a held-out sample of 30%. Performance of models using linked claims and EHR data was compared to models using claims data only.

RESULTS

Among 15 845 047 encounters by 1 574 612 patients, 99 098 (0.6%) were followed by a self-harm event within 90 days. Overall classification performance did not differ between the best-fitting model using all data (area under the receiver operating curve or AUC = 0.846, 95% CI 0.839-0.854) and the best-fitting model limited to data available from insurance claims (AUC = 0.846, 95% CI 0.838-0.853). Competing models showed similar classification performance across a range of cut-points and similar calibration performance across a range of risk strata. Results were similar when the sample was limited to health systems and time periods where PHQ-9 depression questionnaires were recorded more frequently.

CONCLUSION

Investigators using health record data to account for pre-existing risk in observational studies of suicidal behavior need not limit that research to databases including linked EHR data.

Are People Who Die by Intentional Medication Poisoning Dispensed Those Medications in the Year Prior to Death?

OBJECTIVE

The rate of suicidal poisoning in the United States has increased substantially over the past 20 years. Understanding whether prescription medications used for self-poisoning were recently dispensed would help inform suicide prevention efforts. Alternatively, medications for self-poisoning could have been formerly dispensed or collected from friends, family, or illicit sources.

METHODS

Among those who died by intentional opioid and psychotropic poisonings, we conducted a descriptive study to determine what proportion had a recently filled prescription that could have been the means of suicide. Subjects were all people who died by intentional poisoning across nine health-care systems within the NIH-funded Mental Health Research Network.

RESULTS

Among the 3,300 people who died by suicide, 700 died by any poisoning and 194 died by intentional opioid or psychotropic/hypnotic medication poisoning. Among those who died by intentional opioid poisoning 73% were dispensed an opioid in the year prior. Among those who died by intentional psychotropic/hypnotic poisoning, 83% were dispensed any psychotropic and 61% were dispensed a hypnotic in prior year. Most people were continuously dispensed the same medications used in their intentional poisonings in the year prior to death.

CONCLUSIONS

Our results indicate that most medications used in suicidal overdose were likely recently dispensed. Therefore, future suicide prevention studies and prevention resources should focus on medication safety interventions such as lethal-means counseling for medication access, limiting quantities dispensed, opioid antagonists, and blister packs. HIGHLIGHTSUnderstanding whether medications used for self-poisoning were recently dispensed or formerly/never dispensed would help inform future studies and suicide prevention efforts.We found that most people who died by intentional poisoning with opioids or psychotropic/hypnotic medications received frequent dispensings of the medication used for self-poisoning in the year prior to death.Future suicide prevention studies and efforts should focus on medication safety interventions such as lethal-means counseling for medication access, limiting quantities dispensed, opioid antagonists, and blister packs.

Clinical implementation of suicide risk prediction models in healthcare: a qualitative study

BACKGROUND

Suicide risk prediction models derived from electronic health records (EHR) are a novel innovation in suicide prevention but there is little evidence to guide their implementation.

METHODS

In this qualitative study, 30 clinicians and 10 health care administrators were interviewed from one health system anticipating implementation of an automated EHR-derived suicide risk prediction model and two health systems piloting different implementation approaches. Site-tailored interview guides focused on respondents' expectations for and experiences with suicide risk prediction models in clinical practice, and suggestions for improving implementation. Interview prompts and content analysis were guided by Consolidated Framework for Implementation Research (CFIR) constructs.

RESULTS

Administrators and clinicians found use of the suicide risk prediction model and the two implementation approaches acceptable. Clinicians desired opportunities for early buy-in, implementation decision-making, and feedback. They wanted to better understand how this manner of risk identification enhanced existing suicide prevention efforts. They also wanted additional training to understand how the model determined risk, particularly after patients they expected to see identified by the model were not flagged at-risk and patients they did not expect to see identified were. Clinicians were concerned about having enough suicide prevention resources for potentially increased demand and about their personal liability; they wanted clear procedures for situations when they could not reach patients or when patients remained at-risk over a sustained period. Suggestions for making risk model workflows more efficient and less burdensome included consolidating suicide risk information in a dedicated module in the EHR and populating risk assessment scores and text in clinical notes.

CONCLUSION

Health systems considering suicide risk model implementation should engage clinicians early in the process to ensure they understand how risk models estimate risk and add value to existing workflows, clarify clinician role expectations, and summarize risk information in a convenient place in the EHR to support high-quality patient care.

Accuracy of ICD-10-CM encounter diagnoses from health records for identifying self-harm events

OBJECTIVE

Assess the accuracy of ICD-10-CM coding of self-harm injuries and poisonings to identify self-harm events.

MATERIALS AND METHODS

In 7 integrated health systems, records data identified patients reporting frequent suicidal ideation. Records then identified subsequent ICD-10-CM injury and poisoning codes indicating self-harm as well as selected codes in 3 categories where uncoded self-harm events might be found: injuries and poisonings coded as undetermined intent, those coded accidental, and injuries with no coding of intent. For injury and poisoning encounters with diagnoses in those 4 groups, relevant clinical text was extracted from records and assessed by a blinded panel regarding documentation of self-harm intent.

RESULTS

Diagnostic codes selected for review include all codes for self-harm, 43 codes for undetermined intent, 26 codes for accidental intent, and 46 codes for injuries without coding of intent. Clinical text was available for review for 285 events originally coded as self-harm, 85 coded as undetermined intent, 302 coded as accidents, and 438 injury events with no coding of intent. Blinded review of full-text clinical records found documentation of self-harm intent in 254 (89.1%) of those originally coded as self-harm, 24 (28.2%) of those coded as undetermined, 24 (7.9%) of those coded as accidental, and 48 (11.0%) of those without coding of intent.

CONCLUSIONS

Among patients at high risk, nearly 90% of injuries and poisonings with ICD-10-CM coding of self-harm have documentation of self-harm intent. Reliance on ICD-10-CM coding of intent to identify self-harm would fail to include a small proportion of true self-harm events.

Cancer and psychiatric diagnoses in the year preceding suicide

BACKGROUND

Patients with cancer are known to be at increased risk for suicide but little is known about the interaction between cancer and psychiatric diagnoses, another well-documented risk factor.

METHODS

Electronic medical records from nine healthcare systems participating in the Mental Health Research Network were aggregated to form a retrospective case-control study, with ICD-9 codes used to identify diagnoses in the 1 year prior to death by suicide for cases (N = 3330) or matching index date for controls (N = 297,034). Conditional logistic regression was used to assess differences in cancer and psychiatric diagnoses between cases and controls, controlling for sex and age.

RESULTS

Among patients without concurrent psychiatric diagnoses, cancer at disease sites with lower average 5-year survival rates were associated with significantly greater relative risk, while cancer disease sites with survival rates of >70% conferred no increased risk. Patients with most psychiatric diagnoses were at higher risk, however, there was no additional risk conferred to these patients by a concurrent cancer diagnosis.

CONCLUSION

We found no evidence of a synergistic effect between cancer and psychiatric diagnoses. However, cancer patients with a concurrent psychiatric illness remain at the highest relative risk for suicide, regardless of cancer disease site, due to strong independent associations between psychiatric diagnoses and suicide. For patients without a concurrent psychiatric illness, cancer disease sites associated with worse prognoses appeared to confer greater suicide risk.

Validating a predictive algorithm for suicide risk with Alaska Native populations

INTRODUCTION

The American Indian/Alaska Native (AI/AN) suicide rate in Alaska is twice the state rate and four times the U.S. rate. Healthcare systems need innovative methods of suicide risk detection. The Mental Health Research Network (MHRN) developed suicide risk prediction algorithms in a general U.S.

PATIENT POPULATION

METHODS

We applied MHRN predictors and regression coefficients to electronic health records of AI/AN patients aged ≥13 years with behavioral health diagnoses and primary care visits between October 1, 2016, and March 30, 2018. Logistic regression assessed model accuracy for predicting and stratifying risk for suicide attempt within 90 days after a visit. We compared expected to observed risk and assessed model performance characteristics.

RESULTS

10,864 patients made 47,413 primary care visits. Suicide attempt occurred after 589 (1.2%) visits. Visits in the top 5% of predicted risk accounted for 40% of actual attempts. Among visits in the top 0.5% of predicted risk, 25.1% were followed by suicide attempt. The best fitting model had an AUC of 0.826 (95% CI: 0.809-0.843).

CONCLUSIONS

The MHRN model accurately predicted suicide attempts among AI/AN patients. Future work should develop clinical and operational guidance for effective implementation of the model with this population.

Innovations in suicide prevention research (INSPIRE): a protocol for a population-based case-control study

BACKGROUND

Suicide deaths have been increasing for the past 20 years in the USA resulting in 45 979 deaths in 2020, a 29% increase since 1999. Lack of data linkage between entities with potential to implement large suicide prevention initiatives (health insurers, health institutions and corrections) is a barrier to developing an integrated framework for suicide prevention.

OBJECTIVES

Data linkage between death records and several large administrative datasets to (1) estimate associations between risk factors and suicide outcomes, (2) develop predictive algorithms and (3) establish long-term data linkage workflow to ensure ongoing suicide surveillance.

METHODS

We will combine six data sources from North Carolina, the 10th most populous state in the USA, from 2006 onward, including death certificate records, violent deaths reporting system, large private health insurance claims data, Medicaid claims data, University of North Carolina electronic health records and data on justice involved individuals released from incarceration. We will determine the incidence of death from suicide, suicide attempts and ideation in the four subpopulations to establish benchmarks. We will use a nested case-control design with incidence density-matched population-based controls to (1) identify short-term and long-term risk factors associated with suicide attempts and mortality and (2) develop machine learning-based predictive algorithms to identify individuals at risk of suicide deaths.

DISCUSSION

We will address gaps from prior studies by establishing an in-depth linked suicide surveillance system integrating multiple large, comprehensive databases that permit establishment of benchmarks, identification of predictors, evaluation of prevention efforts and establishment of long-term surveillance workflow protocols.

Connecting Research and Practice: Implementation of Suicide Prevention Strategies in Learning Health Care Systems

The health care systems affiliated with the Mental Health Research Network strive to be learning health care systems that identify and address evidence gaps of importance to clinicians, patients, and funders. This column describes how research guides clinical care and clinical care guides research in the area of suicide prevention as well as some of the challenges of conducting embedded research.

Serious Suicide Attempts and Risk of Suicide Death

Background: In the US, more than one million people attempt suicide each year. History of suicide attempt is a significant risk factor for death by suicide; however, there is a paucity of data from the US general population on this relationship. Aim: The objective of this study was to examine suicide attempts needing medical attention as a risk for suicide death. Method: We conducted a case-control study involving eight US healthcare systems. A total of 2,674 individuals who died by suicide from 2000 to 2013 were matched to 267,400 individuals by year and location. Results: Prior suicide attempt associated with a medical visit increases risk for suicide death by 39.1 times, particularly for women (OR = 79.2). However, only 11.3% of suicide deaths were associated with an attempt that required medical attention. The association was the strongest for children 10-14 years old (OR = 98.0). Most suicide attempts were recorded during the 20-week period prior to death. Limitations: Our study is limited to suicide attempts for which individuals sought medical care. Conclusion: In the US, prior suicide attempt is associated with an increased risk of suicide death; the risk is high especially during the period immediately following a nonlethal attempt.

Evaluation of Electronic Health Record-Based Suicide Risk Prediction Models on Contemporary Data

BACKGROUND

Suicide risk prediction models have been developed by using information from patients' electronic health records (EHR), but the time elapsed between model development and health system implementation is often substantial. Temporal changes in health systems and EHR coding practices necessitate the evaluation of such models in more contemporary data.

OBJECTIVES

A set of published suicide risk prediction models developed by using EHR data from 2009 to 2015 across seven health systems reported c-statistics of 0.85 for suicide attempt and 0.83 to 0.86 for suicide death. Our objective was to evaluate these models' performance with contemporary data (2014-2017) from these systems.

METHODS

We evaluated performance using mental health visits (6,832,439 to mental health specialty providers and 3,987,078 to general medical providers) from 2014 to 2017 made by 1,799,765 patients aged 13+ across the health systems. No visits in our evaluation were used in the previous model development. Outcomes were suicide attempt (health system records) and suicide death (state death certificates) within 90 days following a visit. We assessed calibration and computed c-statistics with 95% confidence intervals (CI) and cut-point specific estimates of sensitivity, specificity, and positive/negative predictive value.

RESULTS

Models were well calibrated; 46% of suicide attempts and 35% of suicide deaths in the mental health specialty sample were preceded by a visit (within 90 days) with a risk score in the top 5%. In the general medical sample, 53% of attempts and 35% of deaths were preceded by such a visit. Among these two samples, respectively, c-statistics were 0.862 (95% CI: 0.860-0.864) and 0.864 (95% CI: 0.860-0.869) for suicide attempt, and 0.806 (95% CI: 0.790-0.822) and 0.804 (95% CI: 0.782-0.829) for suicide death.

CONCLUSION

Performance of the risk prediction models in this contemporary sample was similar to historical estimates for suicide attempt but modestly lower for suicide death. These published models can inform clinical practice and patient care today.

Detecting and Assessing Suicide Ideation During the COVID-19 Pandemic

Background

The COVID-19 pandemic prompted a rapid shift to virtual (video and telephone) delivery of mental health care, disrupting established processes for identifying people at increased risk of suicidal behavior.

Methods

Following the shift to virtual care, Kaiser Permanente Washington implemented a series of workflow changes to administer standard screening and monitoring questionnaires at virtual visits and to complete structured suicide risk assessments for patients reporting frequent suicidal ideation. These new workflows included automated distribution of questionnaires via the electronic health record (EHR) patient portal and automated alerts to clinicians regarding indicators of high risk.

Results

In March 2020, in-person mental health visits were rapidly and completely replaced by video and telephone visits. The proportion of mental health visits with completed screening and monitoring questionnaires fell from approximately 80% in early 2020 to approximately 30% in late March, then gradually recovered to approximately 60% by the end of 2020. Among patients reporting frequent suicidal ideation on monitoring questionnaires, the proportion with a recorded suicide risk assessment fell from over 90% in early 2020 to approximately 40% in late March, then gradually recovered to nearly 100% by the end of 2020.

Conclusion

Use of EHR patient portal messaging capabilities can facilitate systematic identification and assessment of suicide risk for patients receiving mental health care by telephone or video visit.

Suicide Behavior Following PHQ-9 Screening Among Individuals With Substance Use Disorders

OBJECTIVES

Individuals with substance use disorders (SUD) are at risk for suicide, but no studies have assessed whether routinely administered screeners for suicidal ideation accurately identify outpatients with SUD who are at risk for suicide attempt or death.

METHODS

Data from more than 186,000 visits by over 55,000 patients with mental health and SUD diagnoses receiving care in 7 health systems were analyzed to determine whether responses to item 9 of the 9-item Patient Health Questionnaire, which assesses frequency of thoughts of death and self-harm, are associated with suicide outcomes after an outpatient visit. Odds of suicide attempt or death were computed using generalized estimating equations.

RESULTS

In bivariate analyses, a nearly 5-fold risk was observed for patients answering "nearly every day" relative to "not at all" among individuals who made a suicide attempt within 90 days (4.9% vs 1.1%; χ2 = 1151, P < 0.0001). At nearly half of visits (46%) followed by a suicide attempt within 90 days, patients responded "not at all." In logistic models, compared to "not at all," all other responses were associated with higher odds of suicide attempt or death within 90 days. Fully adjusted models attenuated results but odds of suicide attempt (AOR = 3.24, CI: 2.69-3.91) and suicide death (AOR = 5.67, CI: 2.0-16.1) remained high for those reporting "nearly every day."

CONCLUSIONS

In people with SUD, increasing Patient Health Questionnaire item 9 response predicts increased risk of subsequent suicidal behavior and should prompt intervention. However, clinicians should realize that those reporting "not at all" are not immune from subsequent suicide risk.

Estimating the association between mental health disorders and suicide: a review of common sources of bias and challenges and opportunities for US-based research

Purpose of review

The purpose of this review is to 1) illuminate prevalent methodological approaches and estimates of association between mental health diagnoses and suicide from the meta-analytic literature; 2) discuss key internal and external validity concerns with these estimates; and 3) highlight some of the unique attributes and challenges in US-based suicide research and opportunities to move the evidence base forward.

Recent findings

Globally, there is considerable variability in measures of association between mental health disorders and suicide and a growing debate over methodological approaches to this research. A high suicide incidence makes the US an outlier, and the decentralized nature of US administrative data poses a unique challenge to data linkage that could otherwise advance this research.

Summary

We offer methodological considerations for future research and discuss opportunities made possible by the recent expansion of the US National Violent Death Reporting System to a nationwide registry.

Association between suicide death and concordance with benzodiazepine treatment guidelines for anxiety and sleep disorders

OBJECTIVE

Guidelines for management of anxiety and sleep disorders emphasize antidepressant medications and/or psychotherapy as first/second-line and benzodiazepines as third-line treatments. We evaluated the association between suicide death and concordance with benzodiazepine guidelines.

METHODS

Retrospective case-control study of patients with anxiety and/or sleep disorders from health systems across 8 U.S. states within the Mental Health Research Network. Suicide death cases were matched to controls on year and health system. Appropriate benzodiazepine prescribing defined as: no monotherapy, no long duration, and/or age < 65 years. The association between guideline concordance and suicide death was evaluated, adjusting for diagnostic and treatment covariates.

RESULTS

Sample included 6960 patients with anxiety disorders (2363 filled benzodiazepine) and 6215 with sleep disorders (1237 filled benzodiazepine). Benzodiazepine guideline concordance was associated with reduced odds for suicide in patients with anxiety disorders (OR = 0.611, 95% CI = 0.392-0.953, p = 0.03) and was driven by shorter duration of benzodiazepine use with concomitant psychotherapy or antidepressant medication. The association of benzodiazepine guideline concordance with suicide death did not meet statistical significance in the sleep disorder group (OR = 0.413, 95% CI = 0.154-1.11, p = 0.08).

CONCLUSIONS

We found reduced odds for suicide in those with anxiety disorders who filled benzodiazepines in short-moderate duration with concomitant psychotherapy or antidepressant treatment.

Variation in patterns of health care before suicide: A population case-control study

BACKGROUND

The United States has experienced a significant rise in suicide. As decision makers identify how to address this national concern, healthcare systems have been identified as an optimal location for prevention.

OBJECTIVE

To compare variation in patterns of healthcare use, by health setting, between individuals who died by suicide and the general population.

DESIGN

Case-Control Study.

SETTING

Eight healthcare systems across the United States.

PARTICIPANTS

2674 individuals who died by suicide between 2000 and 2013 along with 267,400 individuals matched on time-period of health plan membership and health system affiliation.

MEASUREMENTS

Healthcare use in the emergency room, inpatient hospital, primary care, and outpatient specialty setting measured using electronic health record data during the 7-, 30-, 60-, 90-, 180-, and 365-day time periods before suicide and matched index date for controls.

RESULTS

Healthcare use was more common across all healthcare settings for individuals who died by suicide. Nearly 30% of individuals had a healthcare visit in the 7-days before suicide (6.5% emergency, 16.3% outpatient specialty, and 9.5% primary care), over half within 30 days, and >90% within 365 days. Those who died by suicide averaged 16.7 healthcare visits during the year. The relative risk of suicide was greatest for individuals who received care in the inpatient setting (aOR = 6.23). There was both a large relative risk (aOR = 3.08) and absolute utilization rate (43.8%) in the emergency room before suicide.

LIMITATIONS

Participant race/ethnicity was not available. The sample did not include uninsured individuals.

CONCLUSIONS

This study provides important data about how care utilization differs for those who die by suicide compared to the general population and can inform decision makers on targeting of suicide prevention activities within health systems.

Improving pragmatic clinical trial design using real-world data

BACKGROUND

Pragmatic clinical trials often use automated data sources such as electronic health records, claims, or registries to identify eligible individuals and collect outcome information. A specific advantage that this automated data collection often yields is having data on potential participants when design decisions are being made. We outline how this data can be used to inform trial design.

METHODS

Our work is motivated by a pragmatic clinical trial evaluating the impact of suicide-prevention outreach interventions on fatal and non-fatal suicide attempts in the 18 months after randomization. We illustrate our recommended approaches for designing pragmatic clinical trials using historical data from the health systems participating in this study. Specifically, we illustrate how electronic health record data can be used to inform the selection of trial eligibility requirements, to estimate the distribution of participant characteristics over the course of the trial, and to conduct power and sample size calculations.

RESULTS

Data from 122,873 people with patient health questionnaire (PHQ) responses, recorded in their electronic health records between 1 July 2010 and 31 March 2012, were used to show that the suicide attempt rate in the 18 months following completion of the questionnaire varies by response to item nine of the PHQ. We estimated that the proportion of individuals with a prior recorded elevated PHQ (i.e. history of suicidal ideation) would decrease from approximately 50% at the beginning of a trial to about 5%, 50 weeks later. Using electronic health record data, we conducted simulations to estimate the power to detect a 25% reduction in suicide attempts. Simulation-based power calculations estimated that randomizing 8000 participants per randomization arm would allow 90% power to detect a 25% reduction in the suicide attempt rate in the intervention arm compared to usual care at an alpha rate of 0.05.

CONCLUSIONS

Historical data can be used to inform the design of pragmatic clinical trials, a strength of trials that use automated data collection for randomizing participants and assessing outcomes. In particular, realistic sample size calculations can be conducted using real-world data from the health systems in which the trial will be conducted. Data-informed trial design should yield more realistic estimates of statistical power and maximize efficiency of trial recruitment.

Self-Reported Suicidal Ideation as a Predictor of Suicidal Behavior Among Outpatients With Diagnoses of Psychotic Disorders

OBJECTIVE

Individuals with psychotic disorders are at high risk of suicidal behavior. The study examined whether response to item 9 of the Patient Health Questionnaire (PHQ-9), which asks about thoughts of death or self-harm, predicts suicidal behavior among outpatients with diagnoses of psychotic disorders.

METHODS

Electronic health records (EHRs) from seven large integrated health systems were used to identify all outpatient visits by adults with a diagnosis of schizophrenia spectrum psychosis or unspecified psychosis from January 1, 2009, to June 30, 2015, during which a PHQ-9 was completed (N=32,982 visits by 5,947 patients). Suicide attempts over the 90 days following each visit were ascertained from EHRs and insurance claims. Suicide deaths were ascertained from state death certificate files.

RESULTS

Risk of suicide attempt within 90 days of an outpatient visit was .8% among patients reporting no thoughts of death or self-harm and 3.5% among those reporting such thoughts "nearly every day." Over 90 days of follow-up, 47% of suicide attempts occurred among those who reported any recent thoughts of death or self-harm at the sampled visit. Also, 59% of attempts occurred among those reporting thoughts of death or self-harm at the index visit or any visit in the prior year. The number of suicide deaths within 90 days (N=10) was too small to accurately assess the relationship between PHQ-9 item 9 response and subsequent suicide death.

CONCLUSIONS

Among outpatients with psychotic disorders, response to item 9 of the PHQ-9 accurately identified those at increased short-term risk of a suicide attempt.

Protecting Life While Preserving Liberty: Ethical Recommendations for Suicide Prevention With Artificial Intelligence

In the United States, suicide increased by 24% in the past 20 years, and suicide risk identification at point-of-care remains a cornerstone of the effort to curb this epidemic (1). As risk identification is difficult because of symptom under-reporting, timing, or lack of screening, healthcare systems rely increasingly on risk scoring and now artificial intelligence (AI) to assess risk. AI remains the science of solving problems and accomplishing tasks, through automated or computational means, that normally require human intelligence. This science is decades-old and includes traditional predictive statistics and machine learning. Only in the last few years has it been applied rigorously in suicide risk prediction and prevention. Applying AI in this context raises significant ethical concern, particularly in balancing beneficence and respecting personal autonomy. To navigate the ethical issues raised by suicide risk prediction, we provide recommendations in three areas-communication, consent, and controls-for both providers and researchers (2).

Predicting Suicide Attempts and Suicide Deaths Following Outpatient Visits Using Electronic Health Records

OBJECTIVE

The authors sought to develop and validate models using electronic health records to predict suicide attempt and suicide death following an outpatient visit.

METHOD

Across seven health systems, 2,960,929 patients age 13 or older (mean age, 46 years; 62% female) made 10,275,853 specialty mental health visits and 9,685,206 primary care visits with mental health diagnoses between Jan. 1, 2009, and June 30, 2015. Health system records and state death certificate data identified suicide attempts (N=24,133) and suicide deaths (N=1,240) over 90 days following each visit. Potential predictors included 313 demographic and clinical characteristics extracted from records for up to 5 years before each visit: prior suicide attempts, mental health and substance use diagnoses, medical diagnoses, psychiatric medications dispensed, inpatient or emergency department care, and routinely administered depression questionnaires. Logistic regression models predicting suicide attempt and death were developed using penalized LASSO (least absolute shrinkage and selection operator) variable selection in a random sample of 65% of the visits and validated in the remaining 35%.

RESULTS

Mental health specialty visits with risk scores in the top 5% accounted for 43% of subsequent suicide attempts and 48% of suicide deaths. Of patients scoring in the top 5%, 5.4% attempted suicide and 0.26% died by suicide within 90 days. C-statistics (equivalent to area under the curve) for prediction of suicide attempt and suicide death were 0.851 (95% CI=0.848, 0.853) and 0.861 (95% CI=0.848, 0.875), respectively. Primary care visits with scores in the top 5% accounted for 48% of subsequent suicide attempts and 43% of suicide deaths. C-statistics for prediction of suicide attempt and suicide death were 0.853 (95% CI=0.849, 0.857) and 0.833 (95% CI=0.813, 0.853), respectively.

CONCLUSIONS

Prediction models incorporating both health record data and responses to self-report questionnaires substantially outperform existing suicide risk prediction tools.

General Medical, Mental Health, and Demographic Risk Factors Associated With Suicide by Firearm Compared With Other Means

OBJECTIVE

Mitigation of suicide risk by reducing access to lethal means, such as firearms and potentially lethal medications, is a highly recommended practice. To better understand groups of patients at risk of suicide in medical settings, the authors compared demographic and clinical risk factors between patients who died by suicide by using firearms or other means with matched patients who did not die by suicide (control group).

METHODS

In a case-control study in 2016 from eight health care systems within the Mental Health Research Network, 2,674 suicide cases from 2010-2013 were matched to a control group (N=267,400). The association between suicide by firearm or other means and medical record information on demographic characteristics, general medical disorders, and mental disorders was assessed.

RESULTS

The odds of having a mental disorder were higher among cases of suicide involving a method other than a firearm. Fourteen general medical disorders were associated with statistically significant (p<.001) greater odds of suicide by firearm, including traumatic brain injury (TBI) (odds ratio [OR]=23.53), epilepsy (OR=3.17), psychogenic pain (OR=2.82), migraine (OR=2.35), and stroke (OR=2.20). Fifteen general medical disorders were associated with statistically significant (p<.001) greater odds of suicide by other means, with particularly high odds for TBI (OR=7.74), epilepsy (OR=3.28), HIV/AIDS (OR=6.03), and migraine (OR=3.17).

CONCLUSIONS

Medical providers should consider targeting suicide risk screening for patients with any mental disorder, TBI, epilepsy, HIV, psychogenic pain, stroke, and migraine. When suicide risk is detected, counseling on reducing access to lethal means should include both firearms and other means for at-risk groups.

The impact of race and ethnicity on rates of return to psychotherapy for depression

BACKGROUND

There are many limitations with the evidence base for the role of race and ethnicity in continuation of psychotherapy for depression.

METHODS

The study sample consisted of 242,765 patients ≥ 18 years old from six healthcare systems in the Mental Health Research Network (MHRN) who had a new episode of psychotherapy treatment for depression between 1/1/2010 and 12/31/2013. Data were from electronic medical records and organized in a Virtual Data Warehouse (VDW). The odds of racial and ethnic minority patients returning for a second psychotherapy visit within 45 days of the initial session were examined using multilevel regression.

RESULTS

The sample was primarily middle aged (68%, 30-64 years old), female (68.5%), and non-Hispanic white (50.7%), had commercial insurance (81.4%), and a low comorbidity burden (68.8% had no major comorbidities). Return rates within 45 days of the first psychotherapy visit were 47.6%. Compared to their non-Hispanic white counterparts, racial and ethnic minority patients were somewhat less likely to return to psychotherapy for a second visit (adjusted odds ratios [aORs] ranged from 0.80 to 0.90). Healthcare system was a much stronger predictor of return rates (aORs ranged from 0.89 to 5.53), while providers accounted for 21.1% of the variance in return rates.

CONCLUSIONS

Provider and healthcare system variation were stronger predictors of patient return to psychotherapy than race and ethnicity. More research is needed to understand why providers and healthcare systems determine psychotherapy return rates for patients of all racial and ethnic groups.

Major Physical Health Conditions and Risk of Suicide

INTRODUCTION

Most individuals make healthcare visits before suicide, but many do not have a diagnosed mental health condition. This study seeks to investigate suicide risk among patients with a range of physical health conditions in a U.S. general population sample and whether risk persists after adjustment for mental health and substance use diagnoses.

METHODS

This study included 2,674 individuals who died by suicide between 2000 and 2013 along with 267,400 controls matched on year and location in a case-control study conducted in 2016 across eight Mental Health Research Network healthcare systems. A total of 19 physical health conditions were identified using diagnostic codes within the healthcare systems' Virtual Data Warehouse, including electronic health record and insurance claims data, during the year before index date.

RESULTS

Seventeen physical health conditions were associated with increased suicide risk after adjustment for age and sex (p<0.001); nine associations persisted after additional adjustment for mental health and substance use diagnoses. Three conditions had a more than twofold increased suicide risk: traumatic brain injury (AOR=8.80, p<0.001); sleep disorders; and HIV/AIDS. Multimorbidity was present in 38% of cases versus 15.5% of controls, and represented nearly a twofold increased risk for suicide.

CONCLUSIONS

Although several individual conditions, for example, traumatic brain injury, were associated with high risk of suicide, nearly all physical health conditions increased suicide risk, even after adjustment for potential confounders. In addition, having multiple physical health conditions increased suicide risk substantially. These data support suicide prevention based on the overall burden of physical health.